[INS-280] Fix Github "repostories" filter does not respect GHES endpoint

pkg/sources/github/github.go

@@ -449,16 +449,7 @@ func (s *Source) Enumerate(ctx context.Context, reporter sources.UnitReporter) e
 		// filteredRepoCache have an entry in the repoInfoCache.
 		for _, repo := range s.filteredRepoCache.Values() {
 			// Extract the repository name from the URL for filtering
-			repoName := repo
-			if strings.Contains(repo, "/") {
-				// Try to extract org/repo name from URL
-				if strings.Contains(repo, "github.com") {
-					parts := strings.Split(repo, "/")
-					if len(parts) >= 2 {
-						repoName = parts[len(parts)-2] + "/" + strings.TrimSuffix(parts[len(parts)-1], ".git")
-					}
-				}
-			}
+			repoName := extractRepoNameFromUrl(repo)
 
 			// Final filter check - only include repositories that pass the filter
 			if s.filteredRepoCache.wantRepo(repoName) {

pkg/sources/github/github_test.go

@@ -489,6 +489,38 @@ func TestNormalizeRepo(t *testing.T) {
 	assert.Contains(t, err.Error(), "no repositories found")
 }
 
+func TestNormalizeRepo_Enterprise(t *testing.T) {
+	tests := []struct {
+		name       string
+		endpoint   string
+		wantResult string
+	}{
+		{
+			name:       "only host",
+			endpoint:   "https://example.com",
+			wantResult: "https://example.com/org/repo.git",
+		},
+		{
+			name:       "host with path",
+			endpoint:   "https://example.com/api/v3",
+			wantResult: "https://example.com/org/repo.git",
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			source := Source{
+				conn: &sourcespb.GitHub{
+					Endpoint: tt.endpoint,
+				},
+			}
+
+			result, err := source.normalizeRepo("org/repo")
+			assert.NoError(t, err)
+			assert.Equal(t, tt.wantResult, result)
+		})
+	}
+}
+
 func TestHandleRateLimit(t *testing.T) {
 	s := initTestSource(&sourcespb.GitHub{Credential: &sourcespb.GitHub_Unauthenticated{}})
 	ctx := context.Background()
@@ -1237,3 +1269,44 @@ func TestFixBothUnitErrAndUnitOKCalled(t *testing.T) {
 	assert.True(t, errCalled)
 	assert.False(t, okCalled)
 }
+
+func TestExtractRepoNameFromURL(t *testing.T) {
+	tests := []struct {
+		name     string
+		url      string
+		expected string
+	}{
+		{
+			name:     "git URL",
+			url:      "https://github.com/org/repo.git",
+			expected: "org/repo",
+		},
+		{
+			name:     "git URL with trailing slash",
+			url:      "https://github.com/org/repo.git/",
+			expected: "org/repo",
+		},
+		{
+			name:     "git URL without .git",
+			url:      "https://github.com/org/repo",
+			expected: "org/repo",
+		},
+		{
+			name:     "git enterprise URL",
+			url:      "https://example-enterprise.com/org/repo.git",
+			expected: "org/repo",
+		},
+		{
+			name:     "just org/repo",
+			url:      "org/repo",
+			expected: "org/repo",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result := extractRepoNameFromUrl(tt.url)
+			assert.Equal(t, tt.expected, result)
+		})
+	}
+}

pkg/sources/github/repo.go

@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"net/url"
 	"regexp"
 	"strings"
 	"sync"
@@ -365,8 +366,30 @@ func (s *Source) normalizeRepo(repo string) (string, error) {
 	// If it's a repository name (contains / but not http), convert to full URL first
 	if strings.Contains(repo, "/") && !regexp.MustCompile(`^[a-z]+://`).MatchString(repo) {
 		fullURL := "https://github.com/" + repo
+		// If using GitHub Enterprise, adjust the URL accordingly
+		if s.conn != nil && s.conn.Endpoint != "" {
+			u, err := url.Parse(s.conn.Endpoint)
+			if err != nil {
+				return "", fmt.Errorf("invalid enterprise endpoint: %w", err)
+			}
+			// we want to remove any path components from the endpoint and just use the host
+			u.Path = "/" + repo
+			fullURL = u.String()
+		}
 		return giturl.NormalizeGithubRepo(fullURL)
 	}
 
 	return "", fmt.Errorf("no repositories found for %s", repo)
 }
+
+// extractRepoNameFromUrl extracts the "owner/repo" name from a GitHub repository URL.
+// Example: http://github.com/owner/repo.git -> owner/repo
+// If an invalid URL is provided, it returns the original string.
+func extractRepoNameFromUrl(repoURL string) string {
+	u, err := url.Parse(repoURL)
+	if err != nil {
+		return repoURL
+	}
+	cleanedPath := strings.Trim(u.Path, "/")
+	return strings.TrimSuffix(cleanedPath, ".git")
+}